feat: add preserveTagNesting option to maintain invalid HTML nesting #295

Author: taoqf

README.md

@@ -86,6 +86,7 @@ Parse the data provided, wrap the result in a new node, and return the root of t
     comment: false,           		// retrieve comments (hurts performance slightly)
     fixNestedATags: false,    		// fix invalid nested <a> HTML tags 
     parseNoneClosedTags: false, 	// close none closed HTML tags instead of removing them 
+    preserveTagNesting: false,		// preserve invalid HTML nesting instead of auto-closing tags (e.g. <p><p>bar</p></p>)
     voidTag: {
       tags: ['area', 'base', 'br', 'col', 'embed', 'hr', 'img', 'input', 'link', 'meta', 'param', 'source', 'track', 'wbr'],	// optional and case insensitive, default value is ['area', 'base', 'br', 'col', 'embed', 'hr', 'img', 'input', 'link', 'meta', 'param', 'source', 'track', 'wbr']
       closingSlash: true	// optional, default false. void tag serialisation, add a final slash <br/>

src/nodes/html.ts

@@ -1029,6 +1029,10 @@ export interface Options {
 	 */
 	fixNestedATags?: boolean;
 	parseNoneClosedTags?: boolean;
+	/**
+	 * When true, preserves invalid HTML nesting (e.g., <p><p>bar</p></p>) instead of auto-closing tags
+	 */
+	preserveTagNesting?: boolean;
 	blockTextElements: {
 		[tag: string]: boolean;
 	};
@@ -1135,7 +1139,7 @@ export function base_parse(data: string, options = {} as Partial<Options>) {
 
 			const parentTagName = currentParent.rawTagName as IRawTagName;
 
-			if (!closingSlash && kElementsClosedByOpening[parentTagName]) {
+			if (!closingSlash && !options.preserveTagNesting && kElementsClosedByOpening[parentTagName]) {
 				if (kElementsClosedByOpening[parentTagName][tagName]) {
 					stack.pop();
 					currentParent = arr_back(stack);
@@ -1216,7 +1220,7 @@ export function base_parse(data: string, options = {} as Partial<Options>) {
 							if (
 								possibleContainer &&
 								possibleContainer.rawTagName &&
-							  possibleContainer.rawTagName.toLowerCase() === closingTag &&
+								possibleContainer.rawTagName.toLowerCase() === closingTag &&
 								!kElementsClosedByClosingExcept[openTag][closingTag]
 							) {
 								// Update range end for closed tag

test/tests/issues/295.js

@@ -0,0 +1,19 @@
+const { parse } = require('@test/test-target');
+
+describe('issue 295 Invalid HTML element nesting', function () {
+	it('Valid HTML nesting kept', function () {
+		// valid HTML - nesting kept
+		const root = parse(`<div>foo<div>bar</div></div>`);
+		root.outerHTML.should.equal('<div>foo<div>bar</div></div>');
+	});
+	it('Valid HTML p nested inside of p', function () {
+		// invalid HTML (p nested inside of p) - nesting changed
+		const root = parse(`<p>foo<p>bar</p></p>`, { preserveTagNesting: true });
+		root.outerHTML.should.equal('<p>foo<p>bar</p></p>');
+	});
+	it('Valid HTML ul nested inside of p', function () {
+		// invalid HTML (ul nested inside of p) - nesting kept
+		const root = parse(`<p>foo<ul><li>bar</li></ul></p>`);
+		root.outerHTML.should.equal('<p>foo<ul><li>bar</li></ul></p>');
+	});
+});